IN VITRO EXAMINATION OF SECONDARY CARIES USING INFRARED PHOTOTHERMAL RADIOMETRY AND MODULATED LUMINESCENCE

Transcription

1 IN VITRO EXAMINATION OF SECONDARY CARIES USING INFRARED PHOTOTHERMAL RADIOMETRY AND MODULATED LUMINESCENCE by Jungho Kim A thesis submitted in conformity with the requirements for the degree of Master of Applied Science Mechanical and Industrial Engineering University of Toronto Copyright by Jungho Kim, 1

2 In vitro Examination of Secondary Caries Using Infrared Photothermal Radiometry and Modulated Luminescence Abstract Jungho Kim Master of Applied Science Mechanical and Industrial Engineering University of Toronto 1 Dental secondary caries is the carious lesion developed around existing restoration margins. Many new technologies have been developed for caries detection purposes, but their performance is unsatisfactory for the specific purpose of secondary caries diagnosis. Therefore, the development of a novel technology to detect secondary caries has been highly necessary. The objective of this research was to investigate the ability of photothermal radiometry and modulated luminescence to detect secondary caries: wall lesions and outer lesions. Changes in experimental PTR-LUM signals due to sequential demineralization on vertical walls of sectioned tooth samples were investigated. Another study was conducted to investigate how two different types of secondary caries, wall lesions and outer lesions, affect the PTR-LUM signals. The studies demonstrated that PTR-LUM is sensitive to progressive demineralization and remineralization on vertical walls of sectioned tooth samples, as well as to the presence of wall lesions and outer lesions developed around composite restorations. ii

3 Acknowledgments The preparation of this thesis would not have been possible without all the valuable inputs and supports from other senior researchers in many aspects. I would like to thank my supervisor, Professor Andreas Mandelis, for giving me such a great research opportunity, and also for all the motivations, inspiration, and supports that he kindly provided. I could learn many valuable aspects of research work and significantly improve my research skills through the experience. I also want to give special thanks to senior researchers in our laboratory: Dr. Raymond Jeon, Dr. Koneswaran Sivagurunathan and Dr. Anna Matvienko for their direct assistance to my research work, and also to other researchers and graduate students in our laboratory. I also would like to thank Dr. Stephen Abrams for his assistance in clinical aspects of the research, and also for his supports to the development of experimental protocol used for the research. I also thank Dr. Bennett T. Amaechi for all his supports in many ways, such as providing treatment solutions and conducting TMR analysis. iii

4 Table of Contents Abstract.... ii Acknowledgments... iii Table of Contents... iv List of Tables... vii List of Figures... viii 1 Literature Review Human tooth structure Dental caries Dental secondary caries Dental Restoration Evaluation techniques of secondary caries Visual inspection Radiography Laser fluorescence technique (DIAGNOdent) Quantitative light-induced fluorescence technique (QLF) Photothermal radiometry and modulated luminescence... 1 Rationale Preliminary Studies Different Gap-size between the Restoration and the Vertical Tooth Wall Experimental Set-up Sample preparation iv

5 3.1.3 Frequency Scan Results and Discussion Various optical material types adjacent to the tooth Sample Preparation Line Scan Results and Discussion Conclusion... 6 Examination of Demineralization on a Tooth Wall Demineralization on an Entire Vertical Wall Sample preparation Measurement Transverse Microradiography (TMR) Results and Discussion Conclusion.... Theoretical study Mathematical Modeling..... Calculation using Matlab software Results and Discussion..... Conclusions Localized spot demineralization / remineralization on the vertical tooth wall Sample preparation and Measurement Results and Discussion Conclusions v

6 5 Examination of Demineralization around Restoration Margins Sample preparation Results and Discussion Conclusions Overall Conclusions Appendices Appendix A: Experimental Data - Demineralization on an Entire Vertical Wall Sample # 3A Sample # 5A Sample # 5B Appendix B: Experimental Data - Localized spot demineralization / remineralization on the vertical tooth wall Sample # Sample # Sample # Appendix C: Experimental Data Examination of Demineralization around Restoration Margins Category 1: Outer lesions Category : Wall lesions Category 3: Combination of outer and wall lesions References or Bibliography... 1 vi

7 List of Tables Table 1. Summary of direct restorative materials [Adapted from Am Dent Assoc ]... 7 Table. Summary of indirect restorative materials (Adapted from Am Dent Assoc )... 9 Table 3. Correlation coefficients between TMR results (mineral loss and lesion depth) and experimental data (PTR amplitude, PTR phase, and LUM amplitude): correlation coefficients that are statistically significant are bolded and italicized Table. Description of symbols used in Equation.... Table 5. Variables that need to be replaced for the optical case from the thermal case of Equation 1 and Table 6. Absorption coefficients and scattering coefficients of sound and carious toot h enamel Table 7. Treatment guideline for accelerated remineralization for the last weeks Table 8. Sample 1 PTR-LUM scan at a fixed position ( µm away from the vertical interface) after each demineralization and remineralization period with ±error shown for each signal channel. Demineralization occurred from day 1 to 1; and remineralization occurred from day 15 to day Table 9. TMR results of 16 samples (results from Sample# S1 and S 18 were not available)... 7 Table 1. Correlation coefficients between PTR-LUM signals and TMR results: Significantly significant correlation values are bolded and italicized vii

8 List of Figures Figure 1. Structures of a human molar tooth showing its major tissue components (Adapted from Fentress 5) Figure. Scanning electron microscopic (SEM) images of (a) three major tissues of human teeth: enamel, dentin, and pulp; (b) longitudinal view of enamel rods (R) and interrods (IR); and (c) cross-sectional view of enamel rods (R) and interrods (IR) (Nanci 3).... Figure 3. (a) Sketch of dentinal tubules running in parallel to each other through the dentin structure from pulp to enamel (Compend Contin Educ Dent 198) ; and (b) scanning electron microscopy (SEM) image of a piece of dentin showing a number of dentinal tubules (University of Oslo) Figure. Dynamic demineralization-remineralization equilibrium at the plaque-enamel interface (Winston and Bhaskar 1998).... Figure 5. (a) Tooth restored with amalgam restorations; (b) cross-section along A-A showing the locations around an existing restoration where outer lesions and wall lesions develop Figure 6. Generation of the PTR signal Figure 7. Comparison of the sensitivity of caries detection methods relative to histological examination (Adapted from Jeon et al. 1) Figure 8. Schematic diagram of experimental setup for combined PTR and LUM: red arrow indicates the direction of the incident laser light focused on the sample tooth; pink arrow indicates the direction of the LUM signal collected with the photo-detector; and yellow arrow indicates the direction of the PTR signal collected with the PTR detector Figure 9. For the frequency scan in section 3.1, the laser beam (excitation / probe) was focused at the interfacial edge on the tooth sample, indicated with a red dot in the figure; and for line scan in section 3., the laser beam (excitation / probe) was moved along the line in the direction of the red arrow shown in the figure.... viii

9 Figure 1. PTR-LUM frequency scan results collected at the interfacial edge on the side surface of the tooth with different gap sizes between the tooth and the composite restoration... Figure 11. PTR-LUM signals collected at the interfacial edge on the side surface of the tooth with different gap sizes between the tooth and the composite restoration: PTR signals were obtained at Hz; LUM signals were obtained at Hz Figure 1. (a) Glass and (b) Mirror, respectively, in contact with the sectioned surface of the tooth sample.... Figure 13. PTR-LUM linescan results collected at the interfacial edge on the side surface of the tooth with different materials of various optical properties placed adjacent to the vertical wall of the half-tooth sample #AAA. Zero position on x-axis represents the interfacial edge of the tooth Figure 1. Sample preparation showing (a) the cross-sectional plane for the cut; (b) the positioning of the half tooth with respect to the LEGO block boundary Figure 15. Development of artificial caries on entire surface of vertical wall using demineralization gel in a treatment cell: (a) treatment cell closed; (b) treatment cell open; (c) samples mounted on LEGO blocks on a fixed support; (d) close-up image of the demineralization gel flat vertical wall interface by means of surface tension; (e) schematic cross sectional description of (d) Figure 16. (a) Vertical tooth wall showing enamel and dentin layer; (b) PTR-LUM scans were conducted on the scanning (side) surface of the tooth where the enamel layer is approximately 1 mm thick: For frequency scan, the laser beam (excitation / probe) was focused at the interfacial edge on the tooth sample, indicated with a red dot in the figure; and for line scan, the laser beam (excitation / probe) was moved along the line in the direction of the red arrow shown in the figure Figure 17. Timeline for each stage of demineralization and PTR-LUM scans Figure 18. PTR-LUM line-scan results obtained after each demineralization period in Figure 18 with sample A. Figure 16 shows the probed cross-section ix

10 Figure 19. Frequency scan - PTR amplitude obtained after each demineralization period with sample A, collected at different measurement locations on the side surface of the tooth: at the edge; and at 1 μm, μm, 5 μm, 1 mm and mm away from the edge... 3 Figure. Frequency scan - PTR phase obtained after each demineralization period with sample A, collected at different measurement locations on the side surface of the tooth: at the edge; and at 1 μm, μm, 5 μm, 1 mm and mm away from the edge Figure 1. Frequency scan - LUM amplitude obtained after each demineralization period with sample A, collected at different measurement locations on the side surface of the tooth: at the edge; and at 1 μm, μm, 5 μm, 1 mm and mm away from the edge Figure. Frequency scan - LUM phase obtained after each demineralization period with sample A, collected at different measurement locations on the side surface of the tooth: at the edge; and at 1 μm, μm, 5 μm, 1 mm and mm away from the edge Figure 3 (a) Densitometric tracing for sample #A; (b) Microradiographic image of sample #A; (c) TMR results of samples from entire wall demineralization protocol Figure. Sample geometry with the Cartesian coordinate axes used for the modeling Figure 5. Diffuse-optical field as a function of distance away from the source fixed at (x, y, z ) = (,.1, ) m... 5 Figure 6 (a) Photothermal field (a) amplitudes and (b) phase as a function distance away from the source fixed at (x, y, z ) = (,.1, ) m and frequency at Hz Figure 7. Change in optical and thermal properties over treatment time for a typical sample in the fluoride-free treatment group. (A) absorption and (B) scattering coefficient. Vertical dashed lines separate demineralization and remineralization treatments. Layer 1 = surface layer; Layer = lesion body (Adapted from Hellen et al. in J. Biophotonics 11)... 9 Figure 8. Experimental line scan LUM amplitudes with Sample A in section.1: Demineralization on an entire vertical wall x

11 Figure 9. Theoretical - Optical Green's function solution calculated using Matlab software with different conditions (i.e. different optical-loss coefficient, A) of vertical wall: A =.9 m, A/, and A/ Figure 3. The mask with a hole-opening attached to the vertical wall before each demineralization/remineralization Figure 31. Veritcal wall of sample #6 after one day of demineralization. A white spot was created at the top right corner indicated with a red arrow Figure 3. Sample 1 PTR-LUM scan at fixed positions after each demineralization and remineralization period. Demineralization occurred from day 1 to 1; and remineralization occurred from day 15 to day Figure 33. Sample 1 - Line scan data obtained after each demineralization and remineralization period. PTR signals were obtained at Hz and LUM signals were obtained at Hz Figure 3. Sample 1. Frequency scan PTR amplitude data obtained after each demineralization and remineralization period at different measurement locations on the side surface of tooth: at the edge; and at 1 μm, μm, 5 μm, 1 mm and mm away from the edge Figure 35. Sample 1. Frequency scan PTR phase data obtained after each demineralization and remineralization period at different measurement locations on the side surface of tooth: at the edge; and at 1 μm, μm, 5 μm, 1 mm and mm away from the edge... 6 Figure 36. Sample 1. Frequency scan LUM amplitude data obtained after each demineralization and remineralization period at different measurement locations on the side surface of tooth: at the edge; and at 1 μm, μm, 5 μm, 1 mm and mm away from the edge Figure 37. Sample 1. Frequency scan LUM phase data obtained after each demineralization and remineralization period at different measurement locations on the side surface of tooth: at the edge; and at 1 μm, μm, 5 μm, 1 mm and mm away from the edge... 6 Figure 38. (a) A whole tooth sample with a cavity (indicated as a red rectangle); and (b) its crosssection - a rectangular cavity was made, and left side of the cavity was demineralized while the right side remained intact. The cavity was then filled with the composite restoration xi

12 Figure 39. Line scan data with sample S5 in Category 1: outer lesions. Tooth enamel surface treated area range: to 1 μm; Composite restoration-in-place range: 1 to 3 μm; tooth enamel surface intact area range: 3 to μm 68 Figure. Line scan data with sample S1 in Category : wall lesions. Tooth enamel surface treated area range: to 1 μm; Composite restoration-in-place range: 1 to 3 μm; tooth enamel surface intact area range: 3 to μm Figure 1. Line scan data with sample S1 in Category 3: outer lesions. Tooth enamel surface treated area range: to 1 μm; Composite restoration-in-place range: 1 to 3 μm; tooth enamel surface intact area range: 3 to μm 7 xii

13 1 1 Literature Review 1.1 Human tooth structure Human teeth consist of four major tissues that include enamel, dentin, cementum, and dental pulp (Figure 1) (Ross et al. ). These four tissues have different compositions, properties, and also different functionalities. Figure 1. Structures of a human molar tooth showing its major tissue components (Adapted from Fentress 5). Enamel is the outermost layer of the anatomic crown of a human tooth, and it is the hardest and most highly mineralized substance of a human body (Ross et al. 6). Enamel consists of approximately 96 percent of mineral, and percent of water and other organic materials (ten Cate 1998). The primary mineral of enamel is hydroxyapatite, which is a crystalline calcium phosphate whose chemical formula is Ca 1 (PO ) 6 (OH) (Staines et al. 1981). Ameloblast cells are responsible for the creation of enamel, and the creation process can be generally divided into two stages: secretory stage and maturation stage (ten Cate 1998). In the secretory stage, ameloblasts release enamel proteins into the surrounding area. They will be partially mineralized by the enzyme alkaline phosphatase (Ross et al. ), and eventually become an enamel rod and interrod. Enamel rods and interrods have the same composition, but they can be distinguished by the orientation of the calcium phosphate crystals as described in Figure.

14 Figure. Scanning electron microscopic (SEM) images of (a) three major tissues of human teeth: enamel, dentin, and pulp; (b) longitudinal view of enamel rods (R) and interrods (IR); and (c) cross-sectional view of enamel rods (R) and interrods (IR) (Nanci 3). In the maturation stage, final mineralization occurs with ameloblasts attributing to transportation of substances used in the process. Ameloblasts are broken down after the maturation stage, and for this reason, enamel cannot be regenerated by any means. However, enamel is not a static tissue since it can undergo mineralization changes through natural demineralization and remineralization processes (Bath_Balogh and Fehrenbach 6). As mentioned above, enamel is a very highly mineralized and very brittle tissue (ten Cate 1998). So dentin being a less mineralized and less brittle tissue gives a support to the outer enamel layer, and thereby, prevents the fracture of enamel upon masticatory forces (Johnson 1999). Dentin is composed of 7 weight-percent of mineral hydroxyapatite, percent of organic materials, and the rest 1 percent of water (ten Cate 1998). The majority of a human tooth is composed of dentin. Dentin extends from the crown to the root. It is covered by enamel on the crown and by cementum on the root, and it surrounds the entire pulp tissue within it. Dentin has microscopic channels within the structure, which is known as dentinal tubules (Figure 3). They run parallel to each other throughout the dentin layer originated from the pulp to the undersurface of enamel or cementum (Ross et al. 6). Dentinal tubules contain fluid and cells known as odontoblasts, and consequently, dentin has a degree of permeability. The formation of dentin is initiated by the odontoblasts of the pulp, and begins before the formation of enamel.

15 3 (a) (b) Figure 3. (a) Sketch of dentinal tubules running in parallel to each other through the dentin structure from pulp to enamel (Compend Contin Educ Dent 198) ; and (b) scanning electron microscopy (SEM) image of a piece of dentin showing a number of dentinal tubules (University of Oslo). Cementum and pulp are the two remaining major constituents of a human tooth along with the enamel and dentin. Cementum covers the root of a tooth. It is composed of ~5 percent of inorganic materials, 33 percent of organic materials, and percent of water (Bath-Balogh and Fehrenbach 6). Pulp is the innermost part of a human tooth filled with soft connective tissues (ten Cate 1998) containing blood vessels and nerves that enter the tooth from hole at the apex of the root (Bradway et al. 199). Pulp is commonly called the nerve of the tooth. 1. Dental caries Dental caries, also known as dental cavities or tooth decay, is caused by specific types of acidproducing bacteria present in oral environment. Two major groups of such bacteria responsible for initiating caries include Streptococcus mutans and Lactobacillus (Hardie 198; Rogers 8). These bacteria reside around the teeth in a mass called plaque which serves as a biofilm covering outside enamel layer (Leeds Dental Institute). These bacteria form lactic acid through fermentation in the presence of fermentable carbohydrates such as sucrose, fructose, and glucose

16 (Holloway PJ and Moore WJ 1983; Rogers 8). The production of acid results in decreasing ph in the mouth (Ross et al. 6). Enamel is a highly mineralized tissue, and the mineral content of enamel is sensitive to the change or ph level in the mouth. Enamel is usually in an equilibrium state of back-and-forth demineralization and remineralization with the surrounding saliva (Figure ). However, the rate of demineralization becomes faster than remineralization when the ph at the enamel surface drops below 5.5 (Dawes 3) resulting in a net loss of mineral contents of enamel. This chemical process of acids dissolving tooth enamel can be expressed as below (Brown and Theodore 3): Ca 1 (PO ) 6 (OH) (s) + 8H + (aq) 1Ca + - (aq) + 6HPO (aq) + H O (l) Figure. Dynamic demineralization-remineralization equilibrium at the plaque-enamel interface (Winston and Bhaskar 1998). The process of creating the small carious lesion due to the net loss of mineral contents can be reversed in its early stage through remineralization by keeping proper oral hygiene and dietary control (Bots et al. ; Ly KA et al. 6; Am Dent Assoc; Brit Nut Found; Dental Association; Eur Acad Paed Dent; Medline). However, if the lesion remains untreated, loss of enough mineral structures of enamel makes the soft organic material left behind disintegrates. It

17 5 may result in formation of a cavity or a hole on the enamel layer. Once that happens, the destroyed enamel layer cannot be regenerated since ameloblast cells that are responsible for the initial production of enamel are no longer present. The destroyed enamel allows the access of acids and bacteria to the dentin, which might result in further development of caries in dentin area. Dentin is more vulnerable to caries than enamel due to its lower mineral content, hence, caries can develop more rapidly in dentin. Moreover, when dentin is exposed due to the destroyed enamel layer, tooth nerves in the pulp tissue also become exposed through dentinal tubules. It can be a reason for tooth pain. The carious lesion needs to be removed as soon as possible to prevent its further development. The carious lesion is removed using dental instruments such as a dental drill and a spoon. Once the carious lesion is removed, the missing part of the tooth structure needs to be restored due to functional and aesthetic reason. Common restorative materials include amalgam, composite resin, glass-ionomer, porcelain and gold. 1.3 Dental secondary caries Dental secondary caries, one type of dental caries, is defined as the carious lesion that occurs around existing restoration margins (Mjör and Toffenetti ; Kidd 1). There are two types of secondary caries: an outer lesion and a wall lesion (Figure 5) (Hals and Kvinnsland 197; Kidd 1976; Dionysopoulos et al. 1996). The outer lesion is caused by the plaque accumulation at the restoration margin, while the wall lesion is considered to be caused by the presence of a gap at the tooth restoration interface (Jorgensen and Wakumoto 1968; Kidd et al. 1995; Özer 1997; Totiam et al. 7; Nassar and González-Cabezas 11). Several studies have demonstrated the presence of such a gap after the placement of any restorative material (Brännström et al. 198; Rigsby et al. 199; Federlin et al. 1998; Irie et el. ; Huang et al. ; Opdam et al. 3; Iwami et al, 5; Piwowarczyk et al. 5).Such a gap can be created at the tooth-restoration interface due to several reasons including polymerization shrinkage of resin composites upon curing as well as weak bonding to dentin. Another reason could be the presence of excessive residual water from the etching and washing procedures, which can result in small blister-like voids trapped along the surface of the hybrid layer (Purk et al. 7; Tay et al. 1996). Many studies support the presence of bacteria in this gap at the tooth-restoration interface (Qvist 198; Gonzáles-Cabezas et al. 1995; Gonzáles-Cabezs et al. 1999; González-Cabezas et al. ;

18 6 Splieth et al. 3; Kermanshahi et al. 1). The presence of these bacteria causes diffusion of hydrogen ions through the gap, and consequently, creates wall lesions. Diagnosis of secondary caries is the principal cause of restoration replacement in general dental practice regardless of types of restorative materials (Mjör 1981; Klausner and Charbeneau 1985; Qvist et al. 199; Mjör and Toffenetti 199; Mjör 1993; Pink et el. 199; Friedl et al. 199; Friedl et al. 1995; Mjör 1997). Secondary caries also causes early failure of restorations (Hickel and Manhart 1). (a) (b) Figure 5. (a) Tooth restored with amalgam restorations; (b) cross-section along A-A showing the locations around an existing restoration where outer lesions and wall lesions develop. 1. Dental Restoration Dental caries or external trauma can result in structural loss of a tooth. This damaged tooth structure is replaced by dental restorations for the purpose of restoring the function, integrity, and morphology of the original tooth. A tooth should be prepared before the placement of restorations. The preparation process typically includes cutting the carious or structurally unsound portions of the tooth with dental burrs to make space for the restorations.

19 7 Dental restorations can be broadly divided into two types: direct restorations and indirect restorations. Direct restorations are done by inserting restorative material directly into the tooth, while indirect restorations, such as crowns or bridges, are fabricated outside of the mouth (Am Dent Assoc ). Table 1 and summarize various aspects of different types of restorative materials used for direct and indirect restorations, respectively. Factors Material General Description Principal Uses Leakage and Secondary Caries Direct Restorative Dental Materials AMALGAM A mixture of mercury and silver alloy powder that forms a hard solid metal filling. Self-hardening at mouth temperature. Dental fillings and heavily loaded back tooth restorations. Leakage is moderate, but recurrent decay is no more prevalent than other materials. COMPOSITES Direct and Indirect A mixture of submicron glass filler and acrylic that forms a solid toothcolored restoration. Selfor lighthardening at mouth temperature. Esthetic dental fillings and veneers. Leakage is low when properly bonded to underlying tooth; recurrent decay depends on maintenance of the toothmaterial bond. GLASS IONOMERS Self-hardening mixture of fluoride containing glass powder and organic acid that forms a solid tooth colored restoration able to release fluoride. Small non-load bearing fillings, cavity liners and cements for crowns/bridges. Leakage is generally low; recurrent decay is comparable to other direct materials, fluoride release may be beneficial for patients at high risk for decay (Gama-Teixeira et al. 7) RESIN- IONOMERS Self or lighthardening mixture of submicron glass filler with fluoride containing glass powder and acrylic resin that forms a solid tooth colored restoration able to release fluoride. Small non-load bearing fillings, cavity liners and cements for crowns/bridges. Leakage is low when properly bonded to the underlying tooth; recurrent decay is comparable to other direct materials, fluoride release may be beneficial for patients at high risk for decay.

20 8 Overall durability Cavity Preparation Considerations Clinical Considerations Resistance to Wear Resistance to Fracture Good to excellent in large loadbearing restorations. Requires removal of tooth structure for adequate retention and thickness of the filling. Tolerant to a wide range of clinical placement conditions, moderately tolerant to the presence of moisture during placement. Highly resistant to wear. Brittle, subject to chipping on filling edges, but good bulk strength in larger high- load restorations. Good in smallto-moderate size restorations. Adhesive bonding permits removing less tooth structure. Moderate to good in non load-bearing restorations poor in load-bearing. Adhesive bonding permits removing less tooth structure. Moderate to good in non load-bearing restorations; poor in loadbearing. Adhesive bonding permits removing less tooth structure. Must be placed in a well-controlled field of operation; very little tolerance to presence of moisture during placement. Moderately resistant, but less so than amalgam. Moderate resistance to fracture in highload restorations. High wear when placed on chewing surfaces. Low resistance to fracture. Biocompatibility Well-tolerated with rare occurrences of allergenic response. Post-Placement Sensitivity Early sensitivity to hot and cold possible. Occurrence of sensitivity highly dependent on ability to adequately bond the restoration to the underlying tooth. Low. Low to moderate resistance to fracture. Occurrence of sensitivity highly dependent on ability to adequately bond the restoration to the underlying tooth.

21 9 Esthetics Relative Cost to Patient Average Number of Visits To Complete Silver or gray metallic color does not mimic tooth color. Generally lower; actual cost of fillings depends on their size. One. Mimics natural tooth color and translucency, but can be subject to staining and discoloration over time. Moderate; actual cost of fillings depends on their size and technique. One for direct fillings; + for indirect inlays, veneers and crowns. Mimics natural tooth color, but lacks natural translucency of enamel. Moderate; actual cost of fillings depends on their size and technique. One. Mimics natural tooth color, but lacks natural translucency of enamel. Moderate; actual cost of fillings depends on their size and technique. One. Table 1. Summary of direct restorative materials [Adapted from Am Dent Assoc ] Factors Materials General Description Principal Uses Leakage and Secondary Caries Indirect Restorative Dental Materials ALL- PORCELAIN (ceramic) Porcelain, ceramic or glasslike fillings and crowns. Inlays, onlays, crowns and aesthetic veneers. Sealing ability depends on materials, underlying tooth structure and procedure used for placement. PORCELAIN Fused to metal Porcelain is fused to an underlying metal structure to provide strength to a filling, crown or bridge. Crowns and fixed bridges. GOLD ALLOYS (high noble) Alloy of gold, copper and other metals resulting in a strong, effective filling, crown or bridge. Inlays, onlays, crowns and fixed bridges. METAL ALLOYS (nonnoble) Alloys of nonnoble metals with silver appearance resulting in high strength crowns and bridges. Crowns, fixed bridges and partial dentures. The commonly used methods used for placement provide a good seal against leakage. The incidence of recurrent decay is similar to other restorative procedures.

22 1 Durability Cavity Preparation Considerations Clinical Considerations Resistance to Wear Resistance to Fracture Brittle, may fracture under heavy biting. Strength depends greatly on quality of bond to underlying tooth structure. Because strength depends on adequate porcelain thickness, it requires more aggressive tooth reduction during Very strong and durable. Stronger restoration than porcelain alone; moderately aggressive tooth reduction is required. High corrosion resistance prevents tarnishing; high strength and toughness resist fracture and wear. The relative high strength of metals in thin sections requires the least amount of healthy tooth structure removal. preparation. These are multiple step procedures requiring highly accurate clinical and laboratory processing. Most restorations require multiple appointments and laboratory fabrication. Highly resistant to wear, but porcelain can rapidly wear opposing teeth if its surface becomes rough. Prone to fracture when placed under tension or on impact. Biocompatibility Well tolerated. Post-Placement Sensitivity Highly resistant to wear, but porcelain can rapidly wear opposing teeth if its surface becomes rough. Porcelain is prone to impact fracture; the metal has high strength. Well tolerated, but some patients may show allergenic sensitivity to base metals. Resistant to wear and gentle to opposing teeth. Resistant to wear and gentle to opposing teeth. Highly resistant to fracture. Well tolerated. Sensitivity, if present, is usually not material specific. Low thermal conductivity reduces the likelihood of discomfort from hot and cold. Well tolerated, but some patients may show allergenic sensitivity to base metals. High thermal conductivity may result in early postplacement discomfort from hot and cold.

23 11 Esthetics Relative Cost to Patient Average Number of Visits To Complete Color and translucency mimic natural tooth appearance. Higher; at least two office visits and laboratory services. Minimum of two; matching esthetics of teeth may require more visits. Porcelain can mimic natural tooth appearance, but metal limits translucency. Higher; at least two office visits and laboratory services. Minimum of two; matching esthetics of teeth may require more visits. Metal colors do not mimic natural teeth. Higher; requires at least two office visits and laboratory services. Minimum of two Table. Summary of indirect restorative materials (Adapted from Am Dent Assoc ) Composite restoration was used throughout this research among those listed in Table 1 and. One reason can be the very wide usage of this particular restoration in general dental practice due to its superior aesthetics over other traditional materials such as amalgam and gold. Composite restorations can be made to resemble the appearance of the original tooth in terms of colour and shape. Other advantages of composite restoration include the fact that they are free of mercury; corrosion resistant; and thermally nonconductive (Boksman et al. 1986). Another reason for choosing the composite restoration for this research work is that secondary caries is apparently more common with this particular material than other materials including amalgam: it occurs more frequently, and also progresses at a substantially faster rate (Leinfelder 1988). Composite restorations experience shrinkage upon light-curing and also through aging. It causes the material to pull away from the tooth allowing leakage, which can result in development of secondary caries. 1.5 Evaluation techniques of secondary caries Visual inspection Visual inspection method is very widely used by most dental practitioners to examine dental caries, including secondary caries. However, secondary caries occurs at the marginal interface

24 1 between a tooth and restorations, and this interface cannot be investigated by direct visual approach. Therefore, general practitioners often look for stains and marginal defects at the interface as an indication of secondary caries. However, stains at the margins of tooth-coloured restorations are difficult to differentiate from secondary caries. Therefore, stains especially in cracks, fissures and pits cannot be a reliable predictor of the presence of secondary caries (Tyas 1991; Kidd and Beighton 1996). Also, a marginal defect itself is insufficient to determine the presence of secondary caries (Pimenta et al. 1995). Hence, the existing restoration is often removed for more accurate diagnosis of secondary caries. However, even though exsting restoration is removed for clearer view, small secondary caries in its early stage cannot be recognized simply because it is not visible. In this sense, visual inspection alone cannot be used as a reliable tool to diagnose dental secondary caries. It can be used as a complementary method supporting other more firm and reliable diagnosis tools Radiography Radiography is also very widely used as a caries detection aid, particularly with respect to invisible or poorly visible areas. The capability of radiography for secondary caries detection was discussed by some studies (Zoellner et al. ; ; 3). The common conclusion drawn from the studies was that radiography cannot be considered as a reliable technique for secondary caries detection. One reason can be that diagnosis of secondary caries using radiography highly relies on how clinicians interpret the radiographic image. It means that there are no firmly set criteria for diagnosis of secondary caries using radiography, but it rather depends on the subjective decision by the observer (Mileman and van den Hout ) Laser fluorescence technique (DIAGNOdent) Laser fluorescence (DIAGNOdent) is a non-destructive technique used for detecting dental caries based on laser fluorescence which can distinguish carious and healthy teeth using dc laser excitation. A 655 nm monochromatic light is emitted from an optical tip, and the same optical tip can also detect the amount of back-scattered fluorescence (Lussi et al. 1999). With 655 nm infrared light, the fluorescing particles have been identified as bacterial protoporphyrins (Köng et

25 13 al. 1998; Buchalla et al. 8). Therefore, the amount of back-scattered fluorescence is theoretically proportional to the amount of bacterial infection. Many studies were conducted to validate the capability of this technique for detecting caries. Some studies found that laser fluorescence technique is more sensitive for the occlusal posterior teeth than traditional diagnostic methods; but its specificity is lower than visual inspection (Lussi et al. ; Bader and Shugars ; Ricketts 5). Furthermore, some studies concluded that the DIAGNOdent was suitable for detecting small superficial lesions, but not for deep dentinal lesions (Alwas- Danowska et al. ; Jeon et al. ). The high reproducibility of DIAGNOdent was claimed to be an outstanding benefit of this device (Lussi et al. 1; Tranaeus et al. ; Lussi et al. 6; Huth et al. 8). But on the other hand, it was found that the device readings can be negatively affected by the presence of stains, polishing pastes, or adjacent filling materials (Francescut and Lussi 3; Lussi and Reich 5; Lussi et al. 6; Hitij and Fidler 8). Having said that, this technique clearly has some pros and cons. Tranæus et al. (5) and Neuhaus et al. (9) concluded that this technique might provide additional quantitative information for the evaluation of caries activity and risk assessment. Also, Bamzahim et al. (; 5) showed that this technique has a potential to detect outer secondary caries around amalgam and tooth-colored restorative materials. Nevertheless, these techniques did not present enough evidence to be recommended as a substitute for traditional techniques including visual inspection and radiography Quantitative light-induced fluorescence technique (QLF) Quantitative light-induced fluorescence technique (QLF) is a non-destructive optical technique for the quantitative measurement of early carious lesions based on the loss of fluorescence in demineralized enamel (Van der Veen and Josselin de Jong ). Some studies have demonstrated that QLF has excellent reliability for the quantification of smooth surface caries (Al-Khateeb et al. 1998; Lagerweij et al. 1999; Tranaeus et al. ). It was also demonstrated that QLF is more sensitive than visual inspection, and yielded double the number of carious sites (Kühnisch et al. 7). Also, QLF was found to be able to assess lesion activity (Meller et al. 6).

26 1 The conventional methodology to assess the severity of demineralization with QLF requires some rather involved computer analyses (de Josselin de Jong et al. 1995). This analysis becomes more complicated when applied to secondary caries. Some studies were conducted to evaluate the capability of QLF for the detection of secondary caries (Hall et al. 1997; González-Cabezas et al. ; Pretty et al. 3). González-Cabezas et al. () suggested a potential application of QLF for detection of demineralization adjacent to resin composite or amalgam restorations. The results of that study demonstrated that QLF has relatively high sensitivity, but not acceptably high enough to be used as a reliable tool to detect secondary caries. Pretty et al. (3) stated in his paper that, QLF can detect demineralization longitudinally adjacent to restorative materials. However, this could only be used in those restorative materials that exhibit a loss in fluorescence greater than that seen in demineralized enamel, e.g. amalgam. This would exclude the composite materials (which demonstrate no loss in fluorescence) and others such as glass ionomers that exhibit fluorescence loss similar to that of demineralized enamel Photothermal radiometry and modulated luminescence A combination of photothermal radiometry (PTR) and modulated luminescence (LUM) has been developed into a caries detection technique. PTR signals are complementary to LUM signals as a direct result of the complementary nature of non-radiative and radiative de-excitation processes, which are responsible for the PTR and LUM signal generation, respectively. PTR is based on the modulated thermal infrared (black-body or Planck radiation) response of a medium resulted from periodic irradiation of a specimen. The optical energy from the excitation laser is absorbed by the specimen and converted into thermal energy, which results in a modulation in the temperature of the sample surface. An Infrared detector is used to measure the variation in thermal emissions caused by the modulated temperature, which constitutes the PTR signal (Figure 6). The PTR signals generated through the above process contain subsurface information of a medium in the form of a spatially damped temperature depth integral. The frequency dependence of the penetration depth of thermal waves enables depth profiling of materials (Munidasa and Mandelis 199). PTR has been applied to the non-destructive evaluation (NDE) of subsurface features in opaque media (Busse and Walther 199). When PTR is applied to an opaque medium such as a hard dental tissue, depth information of the medium

27 15 can be obtained with the incident laser power being transported into the medium in two distinct modes: conductively and radiatively. The conductive component dominates in a near-surface distance (~5-5 μm) controlled by the thermal diffusivity of the turbid medium and the modulation frequency of the laser source intensity (Browns et al. 197), while the radiative component dominates in a considerably deeper region (~several millimeters) proportional to the optical penetration of the diffusely scattered laser optical field (Nicolaides et al. ). Figure 6. Generation of the PTR signal. LUM is based on the optical-to-radiative energy conversion. Absorption of the optical energy from laser source results in excitation of chromophores to a higher-energy state followed by deexcitation to a lower energy state and emission of longer wavelength photons. This longer wavelength signal constitutes LUM signal, which can be detected by a photodetector. However, LUM signals can also include or may be dominated by laser light reflected or scattered by the sample specimen (i.e. back-scattering). This back-scattered light can be filtered out by placing an optical filter in front of the photodetector window. This method enables that only the signal exceeding a certain wavelength can pass the filter before reaching the photodetector. When applied to dental application, LUM frequency responses from enamel exhibit two relaxation life

28 16 times. The longer one (~ms) is a standard hydroxyapatite relaxation lifetime common to all teeth, which is not sensitive to the defects or overall quality of the enamel (Nicolaides et al. ; ). On the other hand, the shorter lifetime (~μs) was found to be sensitive to the quality of enamel (Nicolaides et al. ; ). However, it was also found that variations in LUM emission fluxes and lifetimes between healthy and carious enamel have a limited depth profilometric character (Jeon et al. ). The PTR-LUM technique was found to have specificity and sensitivity substantially better than other conventional caries detection techniques, such as visual inspection, radiography, and DIAGNOdent (Figure 7) (Jeon et al. ). Moreover, it was demonstrated that PTR has a potential to be a reliable noninvasive tool for the detection of early interproximal (i.e. between teeth) caries that cannot be detected by conventional dental x-rays (Jeon et al. 7). PTR-LUM could also be used to detect near-surface or deep subsurface carious lesions and for monitoring enamel thickness (Jeon et al. ); to detect early enamel and root caries (Jeon et al. 8); to quantify sound enamel or dentin (Nicolaides et al. ); to detect deep lesions and near-surface cracks (Jeon et al. 3); and to examine the effects of varying enamel thicknesses, presence of fillings, and stains on the tooth surface (Jeon et al. 3). Figure 7. Comparison of the sensitivity of caries detection methods relative to histological examination (Adapted from Jeon et al. 1).

29 17 Rationale Dental secondary caries is the carious lesion developed between marginal interface of an existing restoration and a tooth. Since secondary caries occurs at subsurface regions that are not externally visible in most cases, a novel nondestructive technology to detect dental secondary caries has been highly necessitated. Traditional detection methods include visual inspection and radiography; however, they are not able to image or detect caries in these areas at early phase of the caries process. Many new technologies have been developed for caries detection purposes, but sensitivity and specificity remain unsatisfactory for the purpose of secondary caries diagnosis. The main objective of this research is to investigate the ability of the combined technique of PTR and LUM to detect and monitor the development of secondary caries: both outer lesions and wall lesions. Some studies were conducted previously in our laboratory to investigate the capability of the PTR-LUM to detect carious lesions developed on near-surface and subsurface (Jeon et al. ; 8). For most of the studies, PTR-LUM signals were measured directly on the demineralized surface. The same method can be used to investigate the outer lesions. However, for the case of wall lesions, PTR-LUM signals have to be measured on the surface adjacent, and perpendicular, to the demineralized surface. Therefore, a new method will be introduced to study this particular case of the wall lesions, which was never conducted previously. Optothermal characteristics of tooth samples will be investigated with PTR, while LUM will be used to investigate the optical characteristics of the samples. The study is divided into two main parts. The first part includes the chapter 3 and where the PTR-LUM signal behaviour with the presence of wall lesions will be investigated through experiments, supported by theoretical studies. In the other part including chapter 5, signal behaviours with the presence of outer lesions and wall lesions will be investigated through experiments. The main hypothesis made for the experiment is that demineralization and remineralization processes will cause changes in the optical and thermal properties of affected areas of tissues. It will introduce a new boundary condition for both thermal- and optical-wave to the medium, and thereby result in changes in both PTR and LUM signals.

30 18 The successful development of this technology is of intense interest to the dental industry. It will prevent unnecessary excavation of existing restorations as a screening method for secondary caries, which will reduce the waste of time and money for both patients and clinicians. Moreover, if secondary caries can be detected at an early stage, simple clinical treatment will help restore the lesion site to its original healthy condition through re-mineralization process of tooth structures, which will result in long-term clinical restoration success. 3 Preliminary Studies Some preliminary studies were conducted to investigate how PTR-LUM signals respond to different conditions of the vertical wall of a tooth sample, and consequently verify the sensitivity level of the PTR-LUM system. In section 3.1, some PTR-LUM scans were conducted to demonstrate that PTR-LUM is sensitive to different gap sizes between the restoration and the vertical wall of a tooth sample. The objective of the section 3. was to study how PTR-LUM signals react when different types of optical materials were placed against the vertical wall of a tooth sample. Different gap sizes and different optical materials should introduce different boundary conditions to the vertical wall of a tooth. 3.1 Different Gap-size between the Restoration and the Vertical Tooth Wall Experimental Set-up The schematic diagram of the experimental setup for combined PTR and LUM monitoring is shown in Figure 8. A semiconductor laser diode with wavelength 66 nm and maximum power 13 mw was used (Opnext, HL655MG). A laser diode controller (Thorlabs, LDCB) triggered by the software lock-in amplifier (National Instruments, NI61) modulated the laser current. The modulated laser beam was focused on the surface of the tooth sample with an aid of a lens (Thorlabs, LMR1) as indicated with a red arrow in Figure 8. The modulated PTR response emerging from the tooth sample was focused by two off-axis paraboloidal mirrors and collected

31 19 at the mid-infrared HgCdZnTe (MCT) detector (VIGO, PVI-TE-5) as indicated with yellow arrows in Figure 8. The LUM response was focused onto a photodetector (Thorlabs, PDA36A) as indicated with a pink arrow in Figure 8. A cut-on colored optical filter (cut-on wavelength 715 nm, Edmund Optics) was placed in front of the photodetector to block unwanted laser light reflected or scattered by the tooth. Collected PTR-LUM signals were digitized and stored using a software lock-in amplifier (National Instruments, NI61) and the Labview software. The same experimental set-up was used for all the other experiments presented in this report. Figure 8. Schematic diagram of experimental setup for combined PTR and LUM: red arrow indicates the direction of the incident laser light focused on the sample tooth; pink arrow indicates the direction of the LUM signal collected with the photo-detector; and yellow arrow indicates the direction of the PTR signal collected with the PTR detector Sample preparation The sample structure used for this set of preliminary studies is shown in Figure 9. An extracted human tooth was vertically sectioned in half using a diamond tipped cutter, and then this half tooth was mounted on a Lego block with an aid of a plastic epoxy. A small structure of a composite restoration was made on a separate Lego block against the sectioned vertical wall of the half tooth sample. The heavily filled large particle composite restoration material Heliomolar

32 (Ivoclar Vivadent) shade A3 was used throughout all the series of experiments presented in this report. The sample was stored in a humid container to prevent it from over-dehydration. Figure 9. For the frequency scan in section 3.1, the laser beam (excitation / probe) was focused at the interfacial edge on the tooth sample, indicated with a red dot in the figure; and for line scan in section 3., the laser beam (excitation / probe) was moved along the line in the direction of the red arrow shown in the figure Frequency Scan Frequency scan was performed on a fixed position on the tooth surface to examine the frequency dependence of PTR and LUM signals from Hz to Hz. The frequency range can be divided into some intervals. Measurements were taken at one frequency starting from Hz, and frequency was automatically increased to a next higher value using Labview software. Four frequency values were used for most of the experiments throughout the study:, 6, 15 and Hz. There was at least 15 seconds of delay time between measurements at each frequency in order to allow the signals to be stabilized upon the change of frequency.

33 Results and Discussion Measurements were made at, and perpendicularly to, the interfacial edge on the side surface of the tooth sample. The tooth sample was placed at a fixed position, and the laser was fixed at one location (indicated with a red dot in Figure 9) on the tooth sample throughout the experiment. The composite restoration block built on a Lego block was placed on a uni-axial micrometer stage. Controlled movement of the composite restoration using the uni-axial micrometer stage made it possible to have a certain value of a gap size between the restoration and the vertical wall of the tooth. Figure 1 shows the results of the frequency scans with varying gap sizes between the tooth and the composite restoration block: in contact, 1-mm gap, -mm gap, and infinite gap. Figure 1 shows that the effect of gap sizes on PTR signal amplitude was minimal; however, the LUM signal amplitude at large ( infinite ) gap size was clearly affected by the gap size. The effect is associated with LUM reflections from the surface of the composite restoration material back into the tooth through the vertical wall. It is for this reason that the absence of the composite restoration material (i.e. infinite gap) results in the minimum LUM amplitude. The effect on PTR amplitude is minimal even at 1-mm gap size due to the attenuation of the thermal-wave signal over that distance. Figure 11 shows the PTR and LUM signals obtained at the interfacial edge on the side surface of the tooth at a fixed frequency with varying gap sizes. PTR signals were obtained at Hz and LUM signals were obtained at Hz so as to optimize signal-to-noise ratio (SNR) for each channel. Measurements were taken at 1μm increments of gap size from to 1 μm, and at 5 μm increments from 1 to 5 μm. Consistently with Figure 1, PTR plots show that after 1 µm only small changes due to thermal-wave confinement effects appear, and it indicates that the gap larger than 1~ µm behaved as a heat insulator. Clinically acceptable gap size is in the range of 119 to 16 µm (Christensen 1966; McLean et al. 1971), which means that the typical gap size in clinical practice should be less than ~15 µm. Therefore, it is a good indication that our PTR as well as LUM is sensitive to the gap size up to the typical gap size. The PTR phase shows sensitivity to the presence of the restoration up to ~ 5 μm gap size, but error bars are relatively large. Compared to Figure 1, the effect of gap size to the PTR signals seems more pronounced at small gap sizes in Figure 11, but is still minimal. However, LUM plots show

34 that further increase in gap size clearly affects the signals up to the relatively large gap size of μm, which is consistent with Figure 1, where LUM amplitude is not affected much by small gap sizes but significantly decreases with large gap sizes. Amplitude (a.u.) Contact ( gap) 1um gap um gap No restoration 1E-3 Frequency Scan with Varying Gap between the tooth and the restoration PTR AMPLITUDE PTR PHASE -6 Contact ( gap) -7 1um gap um gap No restoration Amplitude (a.u.) Contact ( gap) 1um gap.6 um gap No restoration.5 Amplitude (a.u.) LUM AMPLITUDE LUM PHASE -8 Contact ( gap) -1 1um gap -1 um gap No restoration -1 Figure 1. PTR-LUM frequency scan results collected at the interfacial edge on the side surface of the tooth with different gap sizes between the tooth and the composite restoration

35 3 PTR-LUM with Varying Gap between the tooth and the restoration 6 PTR AMPLITUDE LUM AMPLITUDE Gap (um) Gap (um) PTR PHASE Gap (um) LUM PHASE Gap (um) Figure 11. PTR-LUM signals collected at the interfacial edge on the side surface of the tooth with different gap sizes between the tooth and the composite restoration: PTR signals were obtained at Hz; LUM signals were obtained at Hz.

36 3. Various optical material types adjacent to the tooth 3..1 Sample Preparation The same tooth sample and the composite restoration block used in section 3.1 were used for this section as well. In addition, structures with glass and mirror were prepared as shown in Figure 1. Figure 1. (a) Glass and (b) Mirror, respectively, in contact with the sectioned surface of the tooth sample. 3.. Line Scan Line scan measures PTR and LUM signals along a spatial coordinate on the tooth surface at some fixed frequency. Line scan was conducted along a line on the tooth sample indicated as red dashed arrow in Figure 9. The measurement was started from the interfacial edge of the tooth sample and moved in the direction of the red dashed arrow. The micrometer stage, where the tooth sample was placed, was adjusted to move to the next measurement position. After each change of measurement positions, there was 15 seconds of delay time for stabilization of signals.

37 Results and Discussion Line scans were conducted with various materials adjacent to the tooth sample # AAA. Various materials include composite restoration (Figure 9), glass (Figure 1a), and mirror (Figure 1b). Line scan was also conducted with nothing placed against the vertical tooth wall. It was found that there was a signal strength dependence on the type of material placed against the vertical tooth wall for both PTR and LUM. This is consistent with the effect of vertical surface reflectivity on the overall measured LUM signal, Figure 13. The shape of the PTR curves is indicative of the dominance of thermal-wave confinement within the gap, as well as between the laser beam location and the vertical wall. The highest LUM signal occurs with the mirror in place (highest reflectivity) and the lowest with no material beside the tooth (zero reflectivity). In conclusion, Figure 13 indicates PTR and LUM sensitivity levels which would support signal sensitivity to vertical wall demineralization or breakdown along the margins of the restoration.

38 6 6 5 Line-Scan with different materials PTR LUM PTR None Restoration with um gap Restoration in contact Mirror in contact Glass in contact LUM Figure 13. PTR-LUM linescan results collected at the interfacial edge on the side surface of the tooth with different materials of various optical properties placed adjacent to the vertical wall of the half-tooth sample #AAA. Zero position on x-axis represents the interfacial edge of the tooth. 3.3 Conclusion From section 3.1, it was demonstrated that PTR-LUM system is sensitive to different conditions of the vertical wall of the tooth sample caused by changing the gap size between the composite restoration block and the tooth sample. Also in section 3., PTR-LUM was found to be sensitive to different types of optical materials placed against the vertical wall of the tooth sample.

39 7 Examination of Demineralization on a Tooth Wall.1 Demineralization on an Entire Vertical Wall From section 3, PTR-LUM s capability for detection of some changes in optical properties on vertical tooth wall was demonstrated. Therefore, in this section of the study, the ability of PTR- LUM system to detect demineralization on an entire vertical tooth wall was investigated..1.1 Sample preparation Several extracted human teeth, which are free of cracks, stains, brown spots or white spots were collected. They were vertically sectioned in half with a diamond-tipped cutter ensuring that the natural side surfaces remained intact (Figure 1a). Each half tooth was mounted on a LEGO block while the sectioned tooth surface extended 1 mm beyond the plane of the LEGO block sidewall (Figure 1 b). In total, four samples were prepared and numbered as #A, 3A, 5A, and 5B. Samples prepared through the steps described above were stored in a humid container to prevent over-dehydration because it can result in structural damages of dental tissues, such as micro cracks. The samples were demineralized using a standard demineralization gel in order to develop artificial wall lesions on the entire surface of the vertical walls of the samples. The gel consists of an acidified lactic gel containing.1m lactic acid (C 3 H 6 O 3 ) gelled to a thick consistency with 1% hydroxyethyl cellulose (HEC) and adjusted to ph.5 by the addition of.1m NaOH. The samples were mounted on a fixed support in the treatment cell (Figure 15a and 15b) with the sectioned flat surface facing downward (Figure 15c). Demineralization gel was placed under the tooth sample. Initially the tooth did not touch the gel. Next, the gel surface was slightly elevated until it touched the entire surface of the sectioned wall. The gel was applied only on the sectioned surface while leaving all other surfaces of tooth sample intact. It remained in contact with the flat tooth surface aided by surface tension (Figure 15d and 15e).

40 8 (a) (b) Figure 1. Sample preparation showing (a) the cross-sectional plane for the cut; (b) the positioning of the half tooth with respect to the LEGO block boundary. Figure 15. Development of artificial caries on entire surface of vertical wall using demineralization gel in a treatment cell: (a) treatment cell closed; (b) treatment cell open; (c) samples mounted on LEGO blocks on a fixed support; (d) close-up image of the demineralization gel flat vertical wall interface by means of surface tension; (e) schematic cross sectional description of (d).

41 9.1. Measurement Both enamel and dentin are affected when demineralizing the entire sectioned surface of the tooth (see Figure 16 a). Measurement locations were carefully chosen in each sample, since the thickness of enamel varies from enamel to root. The thinnest section is at the cementoenamel junction (CEJ) or where enamel meets the root (Figure 16 a). Moreover, the thickness of enamel varies between samples. In each case, the measurement line was chosen such that the thickness of the enamel layer was approximately 1mm. Before each scanning, each sample was taken out of the humid container, rinsed thoroughly with flowing clean tap water for more than seconds, and excessive water was removed. Then the sample was placed on a three-axis micrometer sample stage, and the laser was turned on and focused onto the sample tooth by adjusting the three-axis micrometer stage. The actual scanning was conducted at least minutes from the point when the laser was turned on because during minutes the tooth sample was dehydrated properly. The same method was used by many previous studies in our laboratory, and it was adhered to because dehydration of a tooth sample affects its optical qualities such as light scattering and fluorescence, as well as thermal properties (Al Khateeb ). In vitro characteristics of this research made it possible to dehydrate the samples for minutes before each PTR-LUM scan, but it is not very realistic in real clinical or in vivo practice. Therefore, further study is required to fully understand the effects of dehydrations to the signal behaviours. Line and frequency scans were conducted at several stages of demineralization: before demineralization; and after demineralization of 1,, 3, 5, 7, 1, and 1 days, as described schematically in the time-table of Figure 17. This protocol was developed with Dr. B. Amaechi from the University of Texas, San Antonio. Procedures for the line and frequency scan were described in previous sections, 3.. and 3.1.3, respectively.

42 3 Figure 16. (a) Vertical tooth wall showing enamel and dentin layer; (b) PTR-LUM scans were conducted on the scanning (side) surface of the tooth where the enamel layer is approximately 1 mm thick: For frequency scan, the laser beam (excitation / probe) was focused at the interfacial edge on the tooth sample, indicated with a red dot in the figure; and for line scan, the laser beam (excitation / probe) was moved along the line in the direction of the red arrow shown in the figure. Figure 17. Timeline for each stage of demineralization and PTR-LUM scans..1.3 Transverse Microradiography (TMR) After the last PTR-LUM measurement following total 1 days of demineralization, all samples were subjected to transverse microradiography (TMR) analysis in order to determine the degree of demineralization occurred for each sample by measuring the mineral loss and lesion depth. The TMR analysis was conducted at the University of Texas Health Science Center at San Antonio under the supervision of Prof. B.T. Amaechi. Mineral loss and lesion depth values were

43 31 determined by averaging several scans over the distance of the thin section taken from the center of the demineralized area corresponding to the area where PTR-LUM measurements were made. The mineral loss was obtained by calculating the difference in volume percent of minerals between sound and demineralized tissue integrated over the lesion depth. The mineral content plateau in deeper regions of the enamel section, representative of sound tissue, was pre-set at the 87 vol% level (de Josselin de Jong et al. 1987). The lesion depth was determined as the distance from the measured sound enamel surface to the location in the lesion where mineral content was 95% of the sound enamel mineral volume..1. Results and Discussion From the study, it was found that both PTR and LUM signals were affected by the presence of the progressive vertical wall lesion. Figure 18 shows the line-scan data obtained after each period of demineralization with sample A. Experimental results of the other three samples exhibited similar trends, which are included in the appendix section at the end of this report (section 7.1). PTR amplitude plots in Figure 18 clearly show that signal amplitudes decreased upon demineralization, although the trend was not monotonic with treatment time. LUM signals also decreased with demineralization. The non-monotonic temporal behavior of the PTR signal line scans is very interesting as it is consistent with other PTR observations in our lab involving demineralization of uncut (whole) dental enamel surfaces using the same gel. It appears that the state of the demineralization of the vertical wall follows similar trends and the PTR signal is controlled by wall demineralization processes, as expected. In Figure 18, note the thermal-wave interference pattern between laser beam position and vertical wall for distances < 1 mm in the untreated sample and the possible destruction of the interference pattern as a result of the loss of interfacial thermal property step sharpness following even one day of demineralization.

44 3 PTR AMPLITUDE Line Scan (Sample A) 7 6 LUM AMPLITUDE PTR PHASE Note: - PTR signals obtained at Hz, LUM signals obtained at Hz LUM PHASE Figure 18. PTR-LUM line-scan results obtained after each demineralization period in Figure 18 with sample A. Figure 16 shows the probed cross-section. Figure 19 show frequency-scan data obtained after each demineralization period with sample A at different measurement locations on the side surface of the tooth: at the edge; and at 1 μm, μm, 5 μm, 1 mm and mm away from the edge. Figure 19,, 1, and shows PTR amplitude, PTR phase, LUM amplitude, and LUM phase, respectively. Similar to line-scan data shown in Figure 18, PTR signals were found to be more sensitive to demineralization than LUM signals. PTR amplitudes decreased and PTR phases increased with more days of demineralization. It is important to note that the effect of vertical surface demineralization on the PTR signal can be measured up to mm away from the interface in both amplitudes and phases and up to 15 Hz or more. This is consistent with strongly back-scattered light following interaction with the vertical wall which contributes to the generation of the photothermal field at locations away from the interface. At f > 15 Hz, the PTR amplitudes

45 33 become less sensitive to the condition of the vertical wall, as expected from the spatially damped nature of thermal waves. However, the phases remarkably retain their sensitivity to the wall condition even for f ~ Hz. On the other hand, LUM signal amplitudes were less sensitive to demineralization and LUM phases were essentially insensitive, both effects being consistent with the line scans of Figure 18.

46 3 Frequency Scan (Sample A) - PTR 1um um mm um mm Figure 19. Frequency scan - PTR amplitude obtained after each demineralization period with sample A, collected at different measurement locations on the side surface of the tooth: at the edge; and at 1 μm, μm, 5 μm, 1 mm and mm away from the edge.

47 Frequency Scan (Sample A) - PTR EDGE 1um um 5Hz mm mm Figure. Frequency scan - PTR phase obtained after each demineralization period with sample A, collected at different measurement locations on the side surface of the tooth: at the edge; and at 1 μm, μm, 5 μm, 1 mm and mm away from the edge.

48 36 Amplitude (mv) 3 Frequency Scan (Sample A) - LUM EDGE um 5um mm 7 mm 5 3 Figure 1. Frequency scan - LUM amplitude obtained after each demineralization period with sample A, collected at different measurement locations on the side surface of the tooth: at the edge; and at 1 μm, μm, 5 μm, 1 mm and mm away from the edge.

49 37 Frequency Scan (Sample A) - LUM PHASE mm mm Figure. Frequency scan - LUM phase obtained after each demineralization period with sample A, collected at different measurement locations on the side surface of the tooth: at the edge; and at 1 μm, μm, 5 μm, 1 mm and mm away from the edge.

50 38 Transverse microradiography (TMR) analysis was conducted with all the four samples to determine the amount of mineral loss and lesion depth created after total 1 days of artificial demineralization. TMR results are shown in the table of Figure 3 (c) below. (a) (b) Sample A 3A 5A 5B Mineral Loss Vol%.mm Lesion Depth μm Figure 3 (a) Densitometric tracing for sample #A; (b) Microradiographic image of sample #A; (c) TMR results of samples from entire wall demineralization protocol (c) Densitometric trace and associated microradiographic image for sample #A are presented in Figure 3 (a) and (b), respectively. Mineral volume profiles in Figure 3 (a) show a thin layer of lower mineral volume close to the edge. Statistical analysis has been conducted to find the correlation coefficients between PTR-LUM signals and TMR results, and also the statistical significance of the correlation coefficient values. Correlation coefficients and their statistical significance were calculated using Pearson s correlation method and Student s T-test method, respectively, using IBM SPSS Statistics 19 software. PTR-LUM signal data at two stages of artificial demineralization were used to define one set of independent variables for the calculation of the correlation coefficient before demineralization and after a total of 1 days of

51 39 demineralization. For PTR-LUM amplitudes, 1 days of demineralization data were divided by the before demineralization data to calculate percentage signal amplitude change for each sample. For PTR phase, 1 days of demineralization data were subtracted from the before demineralization data to calculate phase shift for each sample. These values were compared to the other independent variables that are mineral loss values and lesion depth values in order to calculate correlation coefficients. The computed correlation coefficients are shown in Table 3 below. In Table 3, the correlation coefficients that were found to be statistically significant are bolded and italicized, with showing their p value beside. In table 3, two values of correlation coefficients were found to be statistically significant: PTR amplitude to the mineral loss (r =.91; p <.5), and LUM amplitude to the lesion depth (r = -.998; p <.1). Other four correlation coefficients were not statistically significant, so were not helpful for investigating the correlations between PTR-LUM channels and TMR results. The main reason for this could be small sample numbers. Another reason could be that the samples experienced relatively low degree of demineralization. Previous studies conducted in our laboratory used the tooth samples demineralized using the same demineralization gel (Jeon et al. 8; Hellen 1). The typical mineral loss value of their sample was more than 1 Vol% µm, and lesion depth was deeper than ~6 µm. The values in Figure 3 (c) are much smaller, and it indicates that the samples used for this study experienced much less degree of demineralization. Higher degree of demineralization of the samples could have resulted in more drastic changes of PTR-LUM signals, and it could have resulted in more correlation coefficient values that are statistically significant. Correlation Coefficient PTR Amplitude change PTR Phase shift LUM Amplitude change Mineral Loss.91 (p <.5) Lesion Depth (p <.1) Table 3. Correlation coefficients between TMR results (mineral loss and lesion depth) and experimental data (PTR amplitude, PTR phase, and LUM amplitude): correlation coefficients that are statistically significant are bolded and italicized.

52 .1.5 Conclusion It was demonstrated that PTR-LUM are capable of detecting the presence and degree of vertical entire wall demineralization. PTR signal amplitudes decreased and phases increased with days of demineralization for both line and frequency scans. LUM signal amplitudes decreased with days of demineralization for both line and frequency scans, but the effects were not as pronounced as PTR signals. LUM signal phases were essentially insensitive. PTR amplitudes exhibited a correlation coefficient of.91 with mineral loss in the statistical analysis. LUM amplitude signals exhibited a correlation coefficient of with lesion depth. These results show that demineralization affected both PTR and LUM signal behavior of each sample and these methods can be used simultaneously, in turn, to monitor full demineralization caries along the vertical wall adjacent to a filling material.. Theoretical study Theoretical study was carried out to aid the interpretation of major parts of the experimental work. Mathematical models for diffuse optical wave and thermal wave were derived by Prof. A. Mandelis using Green s function for LUM and PTR signals, respectively. These derived analytical expressions were calculated using Gaussian-Legendre quadrature rule with Matlab software...1 Mathematical Modeling The geometries and characteristics of the tooth sample with a vertical wall and the appropriate boundary conditions for a diffuse photon density wave field were mathematically modeled by Prof. A. Mandelis. Since vertically half-sectioned teeth must be modeled for the study, it was necessary to derive the Green s functions in three-dimensional geometries with 9 edges and lateral symmetry. The geometries of the tooth sample is shown in Figure 3 below, and the thermal wave Green s function equation used for the modeling is shown as Equation 1.

53 1 Figure. Sample geometry with the Cartesian coordinate axes used for the modeling. (Equation 1) with source location at (x, y, z ) and detector location at (x, y, z) The solution for this thermal wave Green s function in Cartesian coordinates with Robin (or third-kind) boundary conditions was found for the use with the three-dimensional problem of invitro teeth, and shown in Equation below. Table defines all the symbols used in Equation.

54 (Equation ) Symbol Description (x, y, z ) Source location in Cartesian Coordinate (x, y, z) h 1, h k 1, k Detector location in Cartesian Coordinate Convection coefficient of xz-plane and yz-plane, respectively Conduction coefficient of xz-plane and yz-plane, respectively σ Thermal wave number = (1+i)(w/(α)) 1/ w α Angular frequency Thermal diffusivity of enamel µ, η Variable of integration K, K 1 Modified Bessel functions: exponentially decaying function Table. Description of symbols used in Equation.

55 3 The solution for the diffuse optical wave Green s function has exactly the same format as the one for the thermal wave Green s function shown in Equation, except that some variables should be replaced. For the optical case, thermal diffusivity (α) in the thermal case should be replaced by effective optical diffusion coefficient (D eff ); thermal wave number (σ) should be replaced by diffuse-photon-density wave number (σ d ); and (k/h) should be replaced by the optical-loss coefficient (A). Formula for each of these variables is listed in Table 5. The solution of the diffuse optical wave Green s function and thermal wave Green s function involve double integrals of real functions and complex functions, respectively. Variable Formula Note D eff σ d A D eff = vd [m /s] σ d = sqrt(μ α /D) = sqrt(3μ t 'μ α ) A D[(1+r)/(1-r)] For sound enamel: v = speed of light in tooth enamel = 1.93E+8 [m/s] (Meng et al. 9); D = optical diffusion coefficient = (1/3μ t ') [m]; μ t ' = reduced attenuation coefficient = μ α + (1-g)μ s [m -1 ]; μ α = absorption coefficient of enamel = 1 [m -1 ] (Fried et al. 1995); μ s = scattering coefficient of enamel = 6 [m -1 ] (Fried et al. 1995); g = mean cosine of scattering angle =.96 (Fried et al. 1995); r = internal reflection of uniformly diffusing radiation =.65 (Anderson et al. 1989) Table 5. Variables that need to be replaced for the optical case from the thermal case of Equation 1 and... Calculation using Matlab software Computational tools were set up using mathematical computer software, Matlab. Numerical integration of the expressions was performed using Gauss-Legendre quadrature rule. Gaussian quadrature rule calculates an approximation of a definite integral by calculating a weighted sum of its integrand function at specified points within the domain of integration. However, the derived expression in Equation includes some indefinite integrals with upper integration limit being infinity. Therefore, it was necessary to somehow convert these indefinite integrals into the form of definite integrals having a definite upper limit. Equation, which is the complete equation of the thermal wave GF solution, consists of some double integrals of the form of Equation 3. Calculation of Equation 3 is briefly explained for clear understanding of how such an expression was calculated.

56 (Equation 3) Equation 3 consists a double integral, so the inner integral has to be calculated first in order to calculate the outer integral. In order to solve Equation 3, the behavior of the integrand function inside the inner integral was first investigated. The integrand function of the inner integral consists of a decreasing exponential function multiplied by another decreasing Bessel function. This makes the integrand of the inner integral a decreasing function of η. The integrand function of the inner integral was plotted versus η, and the upper limit of the inner integral was determined as where the corresponding integrand remains the same up to at least the third significant digit. Using this definite upper limit, the inner integral was solved using the Gauss- Legendre quadrature rule. However, for more accurate calculation using the Gauss-Legendre quadrature rule, the domain of the inner integral had to be divided into a certain number of subdivisions. The inner integral was calculated with many different of numbers of subdivisions, and the proper number of the divisions of the integral was determined with the criterion that the integral remains the same up to at least third significant digit. The outer integral was computed using the similar method...3 Results and Discussion The optical wave Green s function solution was calculated, and Figure 5 shows the diffuse optical field on the xy-plane as a function of distance away from the source fixed at a certain location: (x, y, z ) = (,.1, ) [m]. The same Cartesian coordinate of the system shown in Figure in section..1 was used for the calculation, and optical properties of sound tooth enamel listed in Table 5 were used. Figure 5 clearly shows that the optical field amplitudes decrease as the detector moves farther from the fixed source.

57 5 Figure 5. Diffuse-optical field as a function of distance away from the source fixed at (x, y, z ) = (,.1, ) m. The thermal wave Green s function solution was computed, and Figure 6 shows the photothermal field on the xy-plane as a function of distance away from the source fixed at a certain location: (x, y, z ) = (,.1, ) [m]. Frequency of Hz was used. The same Cartesian coordinate of the system was used as the optical, and thermophysical properties of the sound enamel were used: convection coefficient, h1 = h = assumed to be 1 [W/mK]; conduction coefficient, k1 = k =.9 [W/mK] (Braden M. 196; Brown WS et al. 197); and thermal diffusivity, α =.5E-7 [W/mK] (Braden M. 196; Brown WS et al. 197). Figure 6a shows the photothermal amplitudes field and Figure 6b shows the photothermal phase field. Similar to the optical field in Figure 5, photothermal field amplitudes decrease and phase lags increase as the detector moves farther from the fixed source location.

58 6 (a) Figure 6 (a) Photothermal field (a) amplitudes and (b) phase as a function distance away from the source fixed at (x, y, z ) = (,.1, ) m and frequency at Hz. (b)

59 7 Figures 5 and 6 clearly show that both diffuse optical and thermal fields behave as expected, i.e. decrease as a function of distance away from the fixed source location. However, these Green function solutions are based on Dirac delta-function sources. Therefore, in order to deal with more realistic Gaussian profile laser beam source, more complicated mathematics is required to find new expressions using these Green function expressions and their spatial integrals. The expressions were derived by Prof. A. Mandelis and are being investigated by another senior researcher in our laboratory, which is beyond the scope of my Master s thesis work. Therefore, for this thesis work, I used the optical wave Green s function solution to approximate the theoretical LUM amplitude signal behavior. The diffusion-wave source generated with the delta function is expected to spread inside the medium while the degree of absorption and scattering is determined by the optical properties of the medium. In the case where the delta-function source is incident on the sample surface and the detector does not coincide with the source (which yields an unbounded value for the Green function), the diffusive delta-function source spreads inside the bulk medium, so a Dirac delta function can adequately approximate a spread focused laser beam when the optical field is sampled at a distance larger than the scattering (or diffusion) length (~ 5-1 mm) away from the source coordinate. Using this method, it was possible to mimic the Gaussian profile beam source, and thereby approximate the theoretical LUM amplitude signal behavior as a function of distance away from the vertical edge. However, theoretical PTR amplitudes and phase couldn t be obtained with the thermal wave Green s function solution using the same method because it was not possible to incorporate the thermal source following non-radiative energy conversion of the optical source because of the deltafunction source required for the Green s function. As shown in Equation, the thermal wave Green s function solution is independent on any optical properties of the medium associated with the absorption and scattering, meaning that it was not possible to have a spread beam source at a certain depth of the medium. Figure 8 shows the experimental data with sample A from section.1: Demineralization on an entire vertical wall. It is essentially the sample plot as the Line scan - LUM amplitude plot in Figure 18, but it shows only three stages of demineralization for the sake of brevity. It shows that LUM amplitudes exhibit a decreasing trend as the sample becomes more demineralized. Figure 9 shows the optical wave Green s function solution as a function of the distance away from the vertical edge with various optical properties of the tooth enamel. The location of the probe and

60 8 the sensor was moved along x-axis down to mm away from the edge while the distance between the two was fixed at 1 µm (in terms of x-axis), and 5 µm (in terms of y-axis i.e. depth). The absorption and scattering coefficients of sound and carious enamel have been obtained from many studies, and listed in Table 6 below. The wavelength used throughout the experiments was 66 nm while the values listed in Table 6 were obtained at 6~633 nm.it indicates the approximate validity of the values in Table 6 to this theoretical study. While the sectioned vertical surface becomes demineralized, the bulk of the medium remains intact. Therefore, for the bulk of the medium, optical properties of sound enamel were used as listed in Table 6. However, these same properties could not be used for the demineralized, sectioned vertical wall. Optical properties of the demineralized vertical wall layer can be characterized by the value of the optical loss coefficient, A (Table 5). Specifically, the coefficient A constitutes the boundary condition of the optical wave Green s function solution, It is inversely proportional to the reduced attenuation coefficient, μ t ', where the reduced attenuation coefficient is basically the weighted sum of absorption coefficient (μ α ) and scattering coefficient (μ s ). Table 6 shows that carious enamel has slightly lower absorption coefficient than sound enamel but has much higher scattering coefficient value, which was also observed in a previous study conducted in our lab (Hellen et al. in J. Biophotonics 11). Figure 7 was adapted from the study by Hellen et al. (Hellen et al. in J. Biophotonics 11) and it clearly shows that absorption coefficient of layer, lesion body, slightly decreases upon demineralization, but scattering coefficient significantly increases upon demineralization.

61 9 Figure 7. Change in optical and thermal properties over treatment time for a typical sample in the fluoride-free treatment group. (A) absorption and (B) scattering coefficient. Vertical dashed lines separate demineralization and remineralization treatments. Layer 1 = surface layer; Layer = lesion body (Adapted from Hellen et al. in J. Biophotonics 11) Therefore, the reduced attenuation coefficient of the carious enamel is generally higher than that of the sound enamel, and hence, the optical-loss coefficient, A is lower for the carious enamel. Figure 9 shows three curves with different values of A: A, A/, and A/5. A was computed as.9 [m] using the optical properties of the sound enamel, and A/ and A/5 are.9/ and.9/5, respectively. Figure 9 shows that the amplitudes are lower for smaller A (more highly scattering medium), as expected, since more light leaks out of the vertical wall and is lost to the diffuse optical field inside the medium. It is concluded that the theoretically predicted decrease of the optical diffusion field strength of the carious enamel (lower A coefficient) than that of sound enamel, Figure 9, behaves consistently with the experimental data (Figure 8). However, the theoretical curves in Figure 9 converge as probe/sensor moves away from the vertical edge, while this behavior wasn t observed with experimental data. The theoretical prediction is expected on physical grounds as at locations distant from the edge the optical-loss boundary effects at the edge become negligible at distances large compared to the scattering length 1/m t. The discrepancy with the experimental results at large distances may be due to the inhomogeneity of optical properties along the surface of the enamel even for slightly different coordinate line scans. The degree of variation of bulk enamel LUM signals shown in Figure 8 is consistent with behavior noticed on other teeth as well. Another possible discrepancy source is the use of the delta function instead of the actual spatially extended optical source of the laser

62 5 beam with the radius of µm. The spatially impulsive source utilized in the simulations is more confined than the near-gaussian semiconductor laser diode beam. The actual spatial distribution of the latter would place part of the optical energy closer to the wall, thereby adding sensitivity to the value of the coefficient A at larger distances away from the wall than the deltafunction distribution after scattering. In addition, the other possible source of discrepancy is that experimental LUM signal may include or even be dominated by back-scattered light from the sample, not the light converted to luminescence. It may be due to inadequate filtering of the incident laser light. Wavelength (nm) Absorption coeff. μ α [m -1 ] Scattering coeff. μ s [m -1 ] References Sound Enamel 63 <1 6 ± 1 8 Fried et al. (1995) 6-55 Groenhuis et al. (1981); Zijp (1) Carious Enamel Spitzer and ten Bosch (1977) Spitzer and ten Bosch (1975) Table 6. Absorption coefficients and scattering coefficients of sound and carious toot h enamel.

63 EXPERIMENTAL DATA LUM AMPLITUDE days demin. 1days demin Figure 8. Experimental line scan LUM amplitudes with Sample A in section.1: Demineralization on an entire vertical wall. Figure 9. Theoretical - Optical Green's function solution calculated using Matlab software with different conditions (i.e. different optical-loss coefficient, A) of vertical wall: A =.9 m, A/, and A/5.

64 5 Theoretical PTR amplitudes and phases simulating the experimental results couldn t be obtained with the thermal-wave Green function solution because, unfortunately, the thermal-wave GF with a thermal spatial impulse is too dissimilar to the broadly spread optical distribution in the bulk of the enamel which is the source of the spatial distribution of the thermal-wave source to which the optical power is converted through local absorption events. However, the foregoing trends in the LUM data with demineralization, and the associated Green function modeling, help the understanding of the PTR trends near and away from the edge. As shown in Figure 19 and in the section.1 of this report: Demineralization on an entire vertical wall, the decrease in PTR amplitude with increasing wall demineralization can, at least partly, be understood in terms of the fact that light escapes across the wall more readily with increased demineralization as indicated by the decreased coefficient A. Therefore, the remaining optical field density in the enamel region decreases, which affects the amount of thermal-wave energy generated upon optical absorption, and the strength (amplitude) of the PTR signal. The trends in the PTR phase cannot be explained through the optical GF behavior as discussed above, however, the non-monotonic behaviors are likely due to thermal-wave confinement between the vertical wall and the center of the thermal-wave source (laser beam location) forming a standing thermal wave trapped within a vertical enamel layer as demonstrated in the study conducted in our lab by Hellen et al. (Hellen et al. in J. Biomed Opt 11)... Conclusions The same trends in optical behavior were observed with the theoretical model calculated with Matlab software as the experimental result. Line scan LUM amplitude signals exhibit a decreasing trend upon demineralization, and the theoretical model showed the same trend. The theoretical model supports the findings from the experimental work in sections.1.

65 53.3 Localized spot demineralization / remineralization on the vertical tooth wall Following the conclusion that PTR-LUM can sense entire-wall demineralization from very early stages, a more clinically relevant localized spot demineralization/remineralization on the vertical tooth wall study was implemented..3.1 Sample preparation and Measurement The sample preparation procedures, all experimental procedures and demineralization protocol were the same as for the entire wall demineralization in section.1, except that only a small circular spot of ~ 1 mm diameter was demineralized. Spot types of lesions were created by using a sticky plastic tape (colored UPVC tape, 3M) as a mask with a punch-through opening where local demineralization occurred upon exposure to a demineralization gel (Figure 3). The mask was applied to the vertical tooth wall before each treatment. After each treatment the mask was removed and the sample was cleaned with flowing tap water in order to remove treatment solution and any residual adhesives. Following the total 1 days of demineralization and PTR- LUM scan, remineralization of the same spot was conducted, and measurements were made, over 6 weeks. For the first weeks, samples were remineralized using artificial saliva that consisted of Ca(NO 3 ) (.35 g/l), KH PO (.15 g/l), KCl (9.69 g/l), and C H 7 AsO (cacodylic acid) (.8 g/l). The ph was adjusted to 7. using KOH. Remineralization treatment each week was started with fresh artificial saliva solution, and during each week of remineralization the solution was replaced after the first three days of remineralization, and after each additional days of further remineralization. Measurements were made on a weekly basis. Artificial saliva alone can remineralize a caries lesion, but not quickly. Therefore, samples were treated with artificial saliva for the first day after the end of the weeks, followed by the more sophisticated remineralization method shown in Table 7 to further accelerate sample remineralization. Measurements were made after the last weeks of remineralization. This method is used to speed up the remineralization of early caries in laboratories testing mouth rinses or toothpastes to simulate the use of these products by human subjects as recommended by general dentists. Demineralization by a demineralizing solution (. mm KH PO, 5 mm acetic acid,. mm of 1M CaCl, and.5 ppm fluoride, ph.5 adjusted by KOH) and remineralization

66 5 by storage in artificial saliva mimics the natural ph cycle in the mouth i.e. during and after eating teeth undergo demineralization; subsequently, saliva remineralizes the demineralized surfaces. The additional remineralization solution (15 ppm NaF) simulates using a fluoride toothpaste or mouth rinse to enhance remineralization, mimicking the use of a therapeutic fluoride product to speed up the remineralization of early caries in the mouth. Figure 3. The mask with a hole-opening attached to the vertical wall before each demineralization/remineralization.

67 55 Treatment duration Treatment type min (started at 9 a.m.) Demineralization solution (ph.5) hours Artificial saliva (ph 7.) 5 min Remineralization solution (15 ppm Fluoride) 3 hours Artificial saliva (ph 7.) 5 min Remineralization solution (15 ppm Fluoride) hours Artificial saliva (ph 7.) min Demineralization solution (ph.5) ~until 9 a.m. next day Artificial saliva (ph 7.) Table 7. Treatment guideline for accelerated remineralization for the last weeks..3. Results and Discussion Figure 31 shows the vertical wall of one sample (sample #6) after one day of demineralization. It is clearly shown that a white spot was created at the top right corner of the vertical wall (indicated with a red arrow in the figure) even after one day of demineralization, which indicates that demineralization actually occurred and made a physical change to the sample. Figure 3 shows PTR-LUM amplitude and phase signals at different measurement locations (at the edge; and at 1 μm, μm, and mm away from the edge) upon progressive demineralization and remineralization with sample 1. Experimental results of the other three samples exhibited similar trends, and are included in the appendix section, section 7.. Figure 3 clearly shows that PTR amplitude decreased upon demineralization. Following remineralization there appeared a delayed upward PTR amplitude trend which was accompanied by a slight downward LUM amplitude trend. The PTR phase exhibited slight increases. It is likely that these trends are due to the changing interface structure at the exposed spot which controls a fraction of the amount of scattered and back-reflected light into the tooth bulk. Table 8 shows the PTR-

68 56 LUM signal values at a fixed position ( µm away from the edge), also depicted in Figure 3. At the bottom of the table, the ± error for each signal channel is also shown. From the table it is clearly seen that even after 1 day of demineralization, the PTR amplitude decreased well beyond the error bar. This makes sense considering that a white spot was created even after one day of demineralization as shown in Figure 31. The PTR phase also changed beyond the error bar only after days of demineralization, but the amount of change with respect to error bar was lower than the PTR amplitude. LUM amplitude changed beyond the error bar after 3 days of demineralization; however, the amount of changes of PTR amplitude after each demineralization was much larger than the one of LUM amplitude. These results indicate that PTR is more sensitive to demineralization than LUM. Signal changes due to remineralization of all channels (excluding LUM phase which was too noisy due to weak signal strength) were generally moderate compared to demineralization, presumably due to the slower rate of remineralization. The PTR amplitude was sensitive only after 1 week of remineralization. PTR phase changed beyond the error bar after 3 weeks of remineralization. LUM amplitude changed beyond the error bar after weeks of remineralization. Figure 33 and Figures 3-37 show line- and frequency-scan data of the same sample, respectively. These figures show that PTR amplitude signals exhibit a slightly decreasing trend in the first 1 days of progressive demineralization and a reversal, a slightly increasing trend, upon the onset of the subsequent remineralization. Also, the PTR phase keeps increasing during both demineralization and remineralization. LUM amplitude signals exhibit a decreasing pattern at excitation / probe distances larger than μm away from the edge for both demineralization and remineralization; however, in locations close to the edge (up to ~ μm), the figures show that LUM signals slightly decrease upon demineralization and slightly increase during the subsequent remineralization. This is consistent with the trends of Figure 3 and implies that probing closer to the vertical surface which undergoes localized demineralization and remineralization enhances LUM and PTR sensitivity to these processes, whereas PTR sensitivity to local conditions on the upper (scanned) surface may control the relative signals at the scanned locations. All PTR frequency scans of Figure 3 and 35 converge above 1 Hz, i.e. they become less sensitive or insensitive to the condition of the vertical wall, as expected when the decreasing thermal diffusion length does not reach the vertical wall (thermally thick condition) for each probe location indicated in the figures. Thermal diffusion length (λ th ) can be calculated

69 57 with the Equation below, and it shows that the thermal diffusion length is reversibly proportional to the frequency. (Eq. ) where α = thermal diffusivity of enamel =.5E-7 [m /s] (Braden M. 196; Brown WS et al. 197); f = laser modulation frequency At 1 Hz, the thermal diffusion length was computed as 37.6 μm using Equation. Figure 3 shows that PTR amplitude frequency scans converge above 1 Hz even at the edge where the thermal diffusion length should be theoretically zero. However, this edge doesn t mean that the center of the beam was at the edge during measurements, but it means that the circumference of the beam profile was at the edge (indicated as red-dot in Figure 9). So at this edge, the center of the beam was actually away from the physical vertical edge of the sample by the amount of the radius of the beam size. Since the radius of the beam is μm, at the edge, the center of the beam was actually ~ μm away from the physical vertical edge of the sample. Therefore, it makes sense that even at the edge, the PTR signals are insensitive to the vertical wall above 1 Hz since the thermal diffusion length cannot reach the vertical wall. Figure 31. Veritcal wall of sample #6 after one day of demineralization. A white spot was created at the top right corner indicated with a red arrow.